Crash while extracting with full_face and --manual-fix

[Kirin-kun]

C:\Users\Kirin>python c:\users\kirin\deepfacelab\main.py extract --input-dir H:\
fakes\harper --output-dir h:\fakes\harper\aligned.ip --detector dlib --face-type
 full_face --manual-fix
Running extractor.

Performing 1st pass...
Running on GeForce GTX 1060 6GB.
100%|██████████████████████████████████████████| 64/64 [00:11<00:00,  5.60it/s]
Performing 2nd pass...
Running on GeForce GTX 1060 6GB.
C:\Program Files\Python36\lib\site-packages\h5py\__init__.py:36: FutureWarning:
Conversion of the second argument of issubdtype from `float` to `np.floating` is
 deprecated. In future, it will be treated as `np.float64 == np.dtype(float).typ
e`.
  from ._conv import register_converters as _register_converters
2018-06-05 20:43:25.681946: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:1356] Found device 0 with properties:
name: GeForce GTX 1060 6GB major: 6 minor: 1 memoryClockRate(GHz): 1.7715
pciBusID: 0000:01:00.0
totalMemory: 6.00GiB freeMemory: 5.58GiB
2018-06-05 20:43:25.690947: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:1435] Adding visible gpu devices: 0
2018-06-05 20:43:27.605056: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:923] Device interconnect StreamExecutor with strength 1
edge matrix:
2018-06-05 20:43:27.611057: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:929]      0
2018-06-05 20:43:27.616057: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:942] 0:   N
2018-06-05 20:43:27.621057: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:1053] Created TensorFlow device (/job:localhost/replica:
0/task:0/device:GPU:0 with 5365 MB memory) -> physical GPU (device: 0, name: GeF
orce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1)
Using TensorFlow backend.
100%|██████████████████████████████████████████| 64/64 [00:12<00:00, 10.21it/s]
Performing manual fix...
Running on GeForce GTX 1060 6GB.
C:\Program Files\Python36\lib\site-packages\h5py\__init__.py:36: FutureWarning:
Conversion of the second argument of issubdtype from `float` to `np.floating` is
 deprecated. In future, it will be treated as `np.float64 == np.dtype(float).typ
e`.
  from ._conv import register_converters as _register_converters
2018-06-05 20:43:57.946792: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:1356] Found device 0 with properties:
name: GeForce GTX 1060 6GB major: 6 minor: 1 memoryClockRate(GHz): 1.7715
pciBusID: 0000:01:00.0
totalMemory: 6.00GiB freeMemory: 5.58GiB
2018-06-05 20:43:57.957792: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:1435] Adding visible gpu devices: 0
2018-06-05 20:43:58.596829: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:923] Device interconnect StreamExecutor with strength 1
edge matrix:
2018-06-05 20:43:58.603829: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:929]      0
2018-06-05 20:43:58.607830: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:942] 0:   N
2018-06-05 20:43:58.612830: I T:\src\github\tensorflow\tensorflow\core\common_ru
ntime\gpu\gpu_device.cc:1053] Created TensorFlow device (/job:localhost/replica:
0/task:0/device:GPU:0 with 5366 MB memory) -> physical GPU (device: 0, name: GeF
orce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1)
Using TensorFlow backend.
 88%|████████████████████████████████████?     | 56/64 [00:46<00:44,  5.53s/it]E
xception while process data [H:\fakes\harper\279209_09big.jpg]: Traceback (most
recent call last):
  File "c:\users\kirin\deepfacelab\utils\SubprocessorBase.py", line 232, in subp
rocess
    result = self.onClientProcessData (data)
  File "c:\users\kirin\deepfacelab\mainscripts\Extractor.py", line 225, in onCli
entProcessData
    landmarks = self.e.extract_from_bgr (image, rects)
  File "c:\users\kirin\deepfacelab\facelib\LandmarksExtractor.py", line 123, in
extract_from_bgr
    image = crop(input_image, center, scale).transpose ( (2,0,1) ).astype(np.flo
at32) / 255.0
  File "c:\users\kirin\deepfacelab\facelib\LandmarksExtractor.py", line 36, in c
rop
    newImg[newY[0] - 1:newY[1], newX[0] - 1:newX[1] ] = image[oldY[0] - 1:oldY[1
], oldX[0] - 1:oldX[1], :]
ValueError: could not broadcast input array from shape (184,0,3) into shape (184
,167,3)

Happened with a face not upright. I had just validated a manual fix and it never processed the next image.

Pic set if you want to test : https://owncloud.dspnet.fr/index.php/s/DFRV1UTnrH2uLmA/download (NSFW)

It doesn't crash without manual-fix, but then it skips the pics where the face is at 90° angle.

[iperov]

fixed